Understanding the Relationship between Inflation and Growth: A Wavelet Transformation Approach in the Case of Bangladesh

• Published date: 01 September 2017
• Date: 01 September 2017
• DOI: http://doi.org/10.1111/twec.12429
• Author: Gazi Salah Uddin,Mohamed Arouri,Bo Sjö,Ahmed Taneem Muzaffar

Understanding the Relationship between

Inflation and Growth: A Wavelet

Transformation Approach in the Case of

Bangladesh

Gazi Salah Uddin

1

, Ahmed Taneem Muzaffar

2

, Mohamed Arouri

3

and Bo Sj€

o

1

1

Department of Management and Engineering, Link€

oping University, Link€

oping, Sweden,

2

School of

Business, University of Western Sydney, Penrith, NSW, Australia and

3

Centre Clermontois de Recherche

en Gestion et Management (CRCGM), Clermont-Ferrand, France

1. INTRODUCTION

WHETHER inflation reduces economic growth or whether some inflation actually stimu-

lates growth is an extensively studied topic in the macroeconomics literature. Ideally,

macroeconomic policies in developing countries should achieve high growth and poverty

reduction with reasonable price stability. The policy issues are when is inflation so high that

it damages growth, are there some optimal inflation that stimulates growth, how much of

inflation is caused by economic growth and what are the effects of inflation on growth over

different timescales?

After several bouts of high inflation in the 1970s, in combination with low growth and

high unemployment, many policymakers tended to believe that keeping inflation at a low

level is a prerequisite for sustained high economic growth. Blanchard and Gal

ı (2007) call this

the ‘divine coincidence’, stabilising the inflation rate at a low level would also stabilise the

output gap. Olivier Blanchard, former chief economist at the International Monetary Fund

(IMF), criticised the view of targeting inflation only and commented that ‘beauty is...not...

synonymous with truth’.

1

In addition, the recent experience with quantitative easing and

government budget austerity around the world has not lifted economies completely out of

recession nor has it caused inflation.

Thus, there is a renewed interest to examine the old debate in recent years. This is partly

because there seems to be a methodological gap in the empirical literature on the growth–in-

flation dynamics. Although the empirical literature investigating the growth–inflation relation-

ship is extensive, covering both developed and developing countries, using single country

data and multicountry panel data, it is common in the literature to focus exclusively on the

time domain. The underlying dynamics in the frequency domain is left out.

However, the dynamic relationship between inflation and growth can differ across different

frequencies. The existing literature fails to reveal some interesting observations that may exist

at different frequencies. For instance, inflation may act as a supply shock in the short run, at

We are grateful to two anonymous referees, Raja Junankar, Ron Ratti, Anis Chowdhury, Stelios Bekiros,

Ramazan Gencßay and Dr. Zhihong Yu, the associate editor of the journal, for their comments and sug-

gestions. An earlier version of the paper appears as Chapter 6 of the second author’s PhD dissertation.

Usual caveats apply.

1

Source: http://www.imf.org/external/np/seminars/eng/2011/res/pdf/ob2presentation.pdf

©2016 John Wiley & Sons Ltd

1918

The World Economy (2017)

doi: 10.1111/twec.12429

The World Economy

high and medium frequencies, and thus affecting output growth. On the other hand, in the

longer run, at lower frequencies, output growth through supply effects may affect inflation. In

short, relying only on the time-domain analysis may provide misleading and incomplete infor-

mation on the causality between inflation and growth. A recent study by Gallegati et al.

(2011, p. 490) state that ‘the “true” economic relationships among variables can be expected

to hold at the disaggregated (scale) level rather than at the usual aggregation level’.

To fill the methodological gap in the literature and to re-examine the old debate on causal-

ity, this paper uses monthly data to investigate the relationship between inflation and growth

in the context of Bangladesh over the period between August 1993 and December 2011 using

a new empirical technique –the wavelet transformation approach. Bangladesh is of interest

because since the early 1990s, it is experiencing an accelerating economic growth and in

comparison with other countries has relatively stable and high inflation.

Instead of examining the time series of growth and inflation time series at their original

level, we first decompose the series into different timescales and investigate the relationship

among components of the decomposed series on a scale-by-scale basis. This enables us to

identify the multiscale causal relationships between inflation and growth, contrary to the tradi-

tional methodologies, such as cointegration, which only distinguish two time frames, that is

short run and long run.

An advantage of using wavelets is that it adjusts its window size optimally to longer basis

functions at low frequencies and to shorter basis functions at high frequencies (Benhmad, 2013).

As a result, it has good frequency resolution for low-frequency movements and good time

resolution for high-frequency movements. This helps overcome the problem associated with the

traditional time-domain approach where a fixed-time window is attached with large numbers of

high and few low-frequency cycles resulting in an overrepresentation of high and an underrepre-

sentation of the low-frequency components. According to Aguiar-Conraria and Soares (2011,

p. 646), a key advantage of wavelet is its ability to perform ‘natural local analysis of a time ser-

ies in the sense that the length of wavelets varies endogenously: it stretches into a long wavelet

function to measure the low-frequency movements; and it compresses into a short wavelet func-

tion to measure the high-frequency movements’. Besides, data on price level and output have a

range of frequencies and can be nonstationary. The wavelet analysis is more suitable in this case

because it does not require a time series to meet the stationary requirement.

After the wavelet transformation of the original series, the co-movement between the infla-

tion and growth can be examined in detail using the cross-wavelet power spectrum and the

cross-wavelet coherence. Next, we perform Granger causality test using the decomposed

wavelets to determine any variation in causality due to differences in frequency and time

domains. In developing countries and from the policymakers’ point of view, the study of the

co-movements between inflation and growth at different frequencies has significant implica-

tions. For instance, given that the developing economies are prone to shocks, the co-move-

ments generated due to certain common shocks that are temporary in nature may require only

partial or no policy response, whereas the co-movements resulting from permanent factors

may require immediate policy response.

The novelty of this study is twofold. First, Bangladesh provides an interesting case to study

macroeconomic issues for developing countries. The country has made commendable progress

in meeting the Millennium Development Goals (MDGs), particularly in the areas of poverty

reduction, ensuring food security, primary school enrolment and infant and maternal mortality

rate, despite political turbulence and natural disasters. A progress report by the Government

of Bangladesh (GOB, 2015) identifies robust GDP growth along with structural transformation

©2016 John Wiley & Sons Ltd

RELATIONSHIP BETWEEN INFLATION AND GROWTH 1919